基于多通道特征融合的病理图像有丝分裂检测
|
摘要
针对有丝分裂核形态多变性而难区分、难检测的问题,提出了一个基于多颜色通道特征融合(MCCFF)的计算机辅助有丝分裂检测算法,主要利用六种颜色通道分别提取特征,再将得到的多通道特征进行串行融合。首先对原病理图像进行去噪,再利用阈值分割将细胞核分割作为候选集;然后对每个候选集细胞核小块分别在不同颜色通道上进行特征提取,将提取出的形态学特征、灰度特征以及纹理特征进行串行融合;最后应用改进的最小距离分类器进行分类。实验以ICPR2012有丝分裂检测大赛的数据集为例,综合评价指标F-measure达到了0.747,结果表明,所提出算法(MCCFF)优于传统算法。
In this paper,mitotic nuclei in breast cancer images are difficult to distinguish and detect because of their morphological variability. In the paper,we propose a computer-assisted mitosis detection algorithm based on multi-color channel feature fusion(MCCFF). Six color channel features were extracted,and obtained multi-channel features were serial fused. Firstly,the noises in the original image were reduced,and then a threshold was used to segment the nucleus to build a candidate set. Then,the candidate nuclei were extracted on different color channels,with the useful morphological,intensity and texture features for fusion. Finally,the improved minimum distance classifier was used to carry out the classification. The experiment takes the data of ICPR2012 mitosis detection competition as an example,and the F-measure reaches 0.747. The results show that the proposed algorithm(MCCFF) is superior to the traditional algorithm.
In this paper,mitotic nuclei in breast cancer images are difficult to distinguish and detect because of their morphological variability. In the paper,we propose a computer-assisted mitosis detection algorithm based on multi-color channel feature fusion(MCCFF). Six color channel features were extracted,and obtained multi-channel features were serial fused. Firstly,the noises in the original image were reduced,and then a threshold was used to segment the nucleus to build a candidate set. Then,the candidate nuclei were extracted on different color channels,with the useful morphological,intensity and texture features for fusion. Finally,the improved minimum distance classifier was used to carry out the classification. The experiment takes the data of ICPR2012 mitosis detection competition as an example,and the F-measure reaches 0.747. The results show that the proposed algorithm(MCCFF) is superior to the traditional algorithm.
引文
[1] C D Malon,C Eric.Classification of mitotic figures with convolutional neural networks and seeded blob features[J].Journal of Pathology Informatics,2013,4(1):9.
[2] K S Beevi,M S Nair,G R Bindu.A Multi-Classifier System for Automatic Mitosis Detection in Breast Histopathology Images Using Deep Belief Networks[J].IEEE Journal of Translational Engineering in Health & Medicine,2017,99(5):1-11.
[3] H Chen,et al.Mitosis detection in breast cancer histology images via deep cascaded networks[C].Thirtieth AAAI Conference on Artificial Intelligence.AAAI Press,2016:1160-1166.
[4] W H ang,et al.Cascaded ensemble of convolutional neural networks and handcrafted features for mitosis detection[J].Proceedings of SPIE-The International Society for Optical Engineering,2014,9041(2):90410B-90410B-10.
[5] K S Beevi,M S Nair,G R Bindu.A Multi-Classifier System for Automatic Mitosis Detection in Breast Histopathology Images Using Deep Belief Networks[J].IEEE Journal of Translational Engineering in Health & Medicine,2017,99(5):1-11.
[6] A Albayrak,G Bilgin.Mitosis detection using convolutional neural network based features[C].IEEE,International Symposium on Computational Intelligence and Informatics.IEEE,2017:335-340.
[7] Dan C Cire?an,A Giusti,M Luca.Gambardella,et al.Mitosis Detection in Breast Cancer Histology Images with Deep Neural Networks[C].MICCAI 2013:Medical Image Computing and Computer-Assisted Intervention-MICCAI,2013:411-418.
[8] F Pourakpour,H Ghassemian.Automated mitosis detection based on combination of effective textural and morphological features from breast cancer histology slide images[C].Biomedical Engineering.IEEE,2016:269-274.
[9] D V K Yarlagadda,P Rao,D Rao.MitosisNet:A Deep Learning Network for Mitosis Detection in Breast Cancer Histopathology Images[C].IEEE International Conference on Biomedical and Health Informatics.IEEE,2017.
[10] B Wu,et al.FF-CNN:An Efficient Deep Neural Network for Mitosis Detection in Breast Cancer Histological Images[J].Communications in Computer and Information Science,June 2017.
[11] H Irshad.Automated mitosis detection in histopathology using morphological and multi-channel statistics features[J].Journal of Pathology Informatics,2013,4(1):10.
[12] A Tashk,et al.An automatic mitosis detection method for breast cancer histopathology slide images based on objective and pixel-wise textural features classification[C].Information and Knowledge Technology.IEEE,2013:406-410.
[13] M Veta,P J V Diest,J P W Pluim.Detecting mitotic figures in breast cancer histopathology images[C].SPIE Medical Imaging,2013:867607.
[14] A M Khan,H El-Daly,N M Rajpoot.A Gamma-Gaussian mixture model for detection of mitotic cells in breast cancer histopathology images[J].Journal of Pathology Informatics,2013,4(4):11.
[2] K S Beevi,M S Nair,G R Bindu.A Multi-Classifier System for Automatic Mitosis Detection in Breast Histopathology Images Using Deep Belief Networks[J].IEEE Journal of Translational Engineering in Health & Medicine,2017,99(5):1-11.
[3] H Chen,et al.Mitosis detection in breast cancer histology images via deep cascaded networks[C].Thirtieth AAAI Conference on Artificial Intelligence.AAAI Press,2016:1160-1166.
[4] W H ang,et al.Cascaded ensemble of convolutional neural networks and handcrafted features for mitosis detection[J].Proceedings of SPIE-The International Society for Optical Engineering,2014,9041(2):90410B-90410B-10.
[5] K S Beevi,M S Nair,G R Bindu.A Multi-Classifier System for Automatic Mitosis Detection in Breast Histopathology Images Using Deep Belief Networks[J].IEEE Journal of Translational Engineering in Health & Medicine,2017,99(5):1-11.
[6] A Albayrak,G Bilgin.Mitosis detection using convolutional neural network based features[C].IEEE,International Symposium on Computational Intelligence and Informatics.IEEE,2017:335-340.
[7] Dan C Cire?an,A Giusti,M Luca.Gambardella,et al.Mitosis Detection in Breast Cancer Histology Images with Deep Neural Networks[C].MICCAI 2013:Medical Image Computing and Computer-Assisted Intervention-MICCAI,2013:411-418.
[8] F Pourakpour,H Ghassemian.Automated mitosis detection based on combination of effective textural and morphological features from breast cancer histology slide images[C].Biomedical Engineering.IEEE,2016:269-274.
[9] D V K Yarlagadda,P Rao,D Rao.MitosisNet:A Deep Learning Network for Mitosis Detection in Breast Cancer Histopathology Images[C].IEEE International Conference on Biomedical and Health Informatics.IEEE,2017.
[10] B Wu,et al.FF-CNN:An Efficient Deep Neural Network for Mitosis Detection in Breast Cancer Histological Images[J].Communications in Computer and Information Science,June 2017.
[11] H Irshad.Automated mitosis detection in histopathology using morphological and multi-channel statistics features[J].Journal of Pathology Informatics,2013,4(1):10.
[12] A Tashk,et al.An automatic mitosis detection method for breast cancer histopathology slide images based on objective and pixel-wise textural features classification[C].Information and Knowledge Technology.IEEE,2013:406-410.
[13] M Veta,P J V Diest,J P W Pluim.Detecting mitotic figures in breast cancer histopathology images[C].SPIE Medical Imaging,2013:867607.
[14] A M Khan,H El-Daly,N M Rajpoot.A Gamma-Gaussian mixture model for detection of mitotic cells in breast cancer histopathology images[J].Journal of Pathology Informatics,2013,4(4):11.